Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:71273
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain ajf.me designates 198.187.29.241 as permitted sender)
Message-ID: <52DC125C.4010609@ajf.me>
Date: Sun, 19 Jan 2014 17:58:52 +0000
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0
MIME-Version: 1.0
To: Sara Golemon <pollita@php.net>, 
 Philip Sturgeon <pjsturgeon@gmail.com>
CC: "internals@lists.php.net" <internals@lists.php.net>
References: <CABDNdAGCH5i9-=4Ys1CXeNaGWGmX=mkTiq0b7BbUKtBPdQ9XVg@mail.gmail.com> <CAESVnVpNm1ektpT=yrE36ZyJrce=Jo-8m3gNEr9Xoz-6uDsibQ@mail.gmail.com>
In-Reply-To: <CAESVnVpNm1ektpT=yrE36ZyJrce=Jo-8m3gNEr9Xoz-6uDsibQ@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] Introducing "Array Of" RFC
From: ajf@ajf.me (Andrea Faulds)



On 16/01/14 13:06, Sara Golemon wrote:
>> https://wiki.php.net/rfc/arrayof
>>
> I'm going to fowl up this conversation by describing what HHVM's Hack
> syntax does for this.  Partly to encourage PHP to do the same, partly
> to just lay out the edge cases we thought about on the way to
> developing the syntax we landed with.
>
> * Generics Syntax - Allows creation of typed classes which repeat
> common functionality with specific implementations
> class Foo<T> {
>    protected T $foo;
>
>    public function getFoo(): T {
>      return $this->foo;
>    }
>
>    pubilc function setFoo(T $foo) {
>      $this->foo = $foo;
>    }
> }
>
> $f = new Foo<Bar>();
> $f->setFoo(new Bar());
> $b = $f->getFoo();
>
> * Typed arrays - Arrays are a special primitive which can be typed as
> implicitly or explicitly numerically indexed, or associative, and to
> specific types:
> array<Tv> or array<Tk,Tv>
>
> function foo(array<Bar> $numIdxArrayOfBar, array<int,Baz>
> $numIdxArrayOfBaz, array<string,$Bong> $assocArrayOfBong) { ... }
>
> Since the value type is unconstrained, it can be a generic type as
> well, so a numerically indexed array of associative arrays of strings
> would look like array<array<string,string>> and so on.
>
> * Nullable types - Any type preceeded by a question mark may be that
> type or null:
> function foo(?Bar $barOrNull, Baz $bar, ?Blong $blongOrNull) {...}
>
> * Soft types - Any type preceeded by an at sign is hinted as that
> type, but not checked:
> function foo(@Bar $IHopeYoureABarObject) {...}
>
> ----------------------------------------------
>
> Within the scope of this RFC, I want to encourage
> array<Tv>/array<Tk,Tv> syntax (with nullable/soft modifiers) as I
> think they cover every edge case presented by this discussion so far.
> It also covers elements not handled by the proposed syntax (index
> types). The syntax will be familiar to developers coming from C++ and
> it leaves room to entertain generics in general.
>
> -Sara
>

After a little thought, I think I'd rather we use the HHVM syntax here. 
Having "Array" for an untyped array but "Bar[]" for a typed one is 
silly. "Array" and "Array<Bar>" make more sense. Being able to type keys 
would be nice, too.

However, I have but one little issue with the HHVM syntax. Why is the ? 
before the type name? I much prefer the C#-style Bar? to this confusing 
?Bar. I suppose this is only a minor issue, though.

-- 
Andrea Faulds
http://ajf.me/